Research map / Vision-language and multimodal / Human and embodied modeling
Diverse Beam Search for Improved Description of Complex Scenes
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace beam search with diverse beam search | captioning LSTM · COCO | Oracle SPICE@20: 27.34 -> 28.09 | Helped |
| replace beam search with diverse beam search | captioning LSTM · COCO | distinct 4-grams: 5.67 -> 13.44 | Helped |
| replace BS with DBS on harder images | captioning LSTM · COCO by image complexity | Oracle SPICE@20 gain: 0.67 / 0.91 / 1.13 | Helped |
| replace BS with DBS | captioning LSTM · PASCAL-50S | Oracle SPICE@20: 8.74 -> 9.293 | Helped |
| replace BS with DBS | VQG model · VQG | SPICE: 15.17 -> 16.49 | Helped |
| replace BS with DBS | GRU NMT · En-Fr newstest-2013 | BLEU-4: 13.52 -> 13.69 | Helped |
| DBS vs Li et al. (2015) MMI decoding | GRU NMT · En-Fr newstest-2013 | Oracle BLEU-4@20: 18.86 -> 18.77 | Hurt |
| replace BS with DBS | captioning LSTM · COCO human study | human preference: 60% | Helped |
| use hamming diversity instead of n-gram, cumulative or embedding diversity (ablation) | DBS · PASCAL-50S | Oracle CIDEr@k | Helped |
| set groups G=B (max exploration) vs G=1 (ablation) | DBS · validation | oracle accuracy | Helped |
Closest papers by meaning
- GLaMM: Pixel Grounding Large Multimodal Model (CVPR 2024 · similarity 0.87)
- Revisiting the Role of Language Priors in Vision-Language Models (ICML 2024 · similarity 0.86)
- GeM-VG: Towards Generalized Multi-image Visual Grounding with Multimodal Large Language Models (AAAI 2026 · similarity 0.86)
- An Empirical Study of Attention and Diversity for Adaptive Visual Token Pruning in Large Vision-Language Models (ICLR 2026 · similarity 0.86)
- Grounding Multimodal Large Language Models to the World (ICLR 2024 · similarity 0.86)
- Multilingual Diversity Improves Vision-Language Representations (NeurIPS 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.